Cloud center energy consumption control for predictability in neural fuzzy systems

Abstract

Cloud computing provides tool-based IT services to customers worldwide. It is built on a fee-for-service model and enables the hosting of large programs from consumer, scientific, and commercial domains. However, cloud application hosting data centers have enormous energy consumption, which leads to high operating expenses and a carbon footprint. Cloud centers get a lot of requests from different users every day. The centers need strong servers that consume a lot of energy and peripherals in order to run in order to meet these needs. It should go without saying that reducing cloud center energy usage requires efficient resource utilization. We have taken a novel hybrid method in this study. Neural fuzzy systems have been employed for the dynamic resource allocation in the cloud with the goal of minimizing energy consumption and load prediction, and the ant colony optimization method has been utilized for the virtual machine's movement. The experimental results show that the proposed strategy works optimally, as compared to other methods in the literature that have the same objective. When compared to other comparable approaches, the suggested solution has an average resource loss of 23.1 % in 720 time periods and an average request denial rate that is 6.5 % lower. Moreover, extensive testing with the SPECpower benchmark has confirmed the proposed methodology, yielding a forecast accuracy of over 97 % for the proposed energy consumption model.

Keywords

• Cloud center
• Neuro-fuzzy
• Classification
• Dynamic allocation of resources
• Cloud computing
• Energy consumption